Yüksel Ayhan, Tombakoğlu Mehmet
Hacettepe University, Nuclear Engineering Department, 06800, Ankara, Turkey.
Hacettepe University, Nuclear Engineering Department, 06800, Ankara, Turkey.
Appl Radiat Isot. 2022 Aug;186:110295. doi: 10.1016/j.apradiso.2022.110295. Epub 2022 May 18.
Detailed geometric information of a high-purity Ge (HPGe) detector is a very important issue for Monte Carlo simulation of the detector. Commonly, users have no geometric information about the detector and information given by the manufacturer is not completely valid for simulation. An equivalent geometry of detector, the parameters of which can be used for Monte Carlo simulation, is optimised using a genetic algorithm for a large-volume HPGe detector in this study. A mixed-point gamma calibration standard, emitting 12 useful gamma-radiation energies within 59.5-1836.1 keV, is placed at 74 different locations around the detector for this purpose. A high-quality solution is generated starting from an initial population of randomly-generated detector geometries using a genetic algorithm. Fitness of each geometry is obtained by comparing full energy peak efficiencies computed by Monte Carlo simulation with experimental values for each energy and position. Efficiencies with relative errors less than 5% for high energies and less than 7% for lower energies, except 59.5 keV, are obtained using optimised equivalent geometry parameters for the Monte Carlo simulation. Also, the necessity of using crystal dimensions smaller than real dimensions for Monte Carlo simulations of high-volume HPGe detectors is discussed. In addition, for Monte Carlo simulation of high-volume HPGe detectors, it is demonstrated that the use of smaller crystal dimensions than the real dimensions is necessary to obtain experimentally measured efficiencies of the detector.
对于高纯锗（HPGe）探测器的蒙特卡罗模拟而言，详细的探测器几何信息是一个非常重要的问题。通常情况下，用户没有探测器的几何信息，而制造商提供的信息对于模拟来说并不完全有效。在本研究中，针对一个大体积HPGe探测器，使用遗传算法优化了一种探测器的等效几何结构，其参数可用于蒙特卡罗模拟。为此，将一个发射59.5 - 1836.1 keV范围内12种有用伽马射线能量的混合点伽马校准标准源放置在探测器周围74个不同位置。从随机生成的探测器几何结构的初始种群开始，使用遗传算法生成高质量的解决方案。通过将蒙特卡罗模拟计算的全能峰效率与每种能量和位置的实验值进行比较，获得每种几何结构的适应度。使用优化后的等效几何结构参数进行蒙特卡罗模拟，对于高能，相对误差小于5%，对于除59.5 keV外的低能，相对误差小于7%时可获得效率。此外，还讨论了在大体积HPGe探测器的蒙特卡罗模拟中使用小于实际尺寸的晶体尺寸的必要性。另外，对于大体积HPGe探测器的蒙特卡罗模拟，结果表明使用小于实际尺寸的晶体尺寸对于获得探测器的实验测量效率是必要的。
